Now, a computer programme can diagnose depression better than psychiatrists
As per the World Health Organization, over 350 million people worldwide suffer from clinical depression. The actual numbers could be much larger, given how easily depression eludes diagnosis. But where the medical fraternity has met with limited success, an algorithm could perhaps be the best bet for an early diagnosis. The Economist reports that a pair of researchers, from the University of South Carolina and Carnegie Mellon University, have developed a computer programme that can, at the moment, diagnose clinical depression with 85% reliability by just recording the subject’s expressions and speech.The algorithm runs facial cues such as the duration of smiles and the length for which the subject looks at the ground in a conversation against templates, which gives a 75% accurate diagnosis. The remaining 10% comes from its ability to quantify observations hitherto thought of as qualitative; for instance, depression patients tend to run their vowels together more often than those who aren’t suffering from the condition while speaking—the algorithm studies its frequency and intensity.
As the programme is made competent for studying more traits and symptoms, the accuracy, there can be little doubt, will reach 100%. Depression is widely acknowledged as a silent killer—over 90% of suicide cases, as per the University of Washington, can be linked to diagnosable mental illnesses like depression. While the software promises faster and more accurate diagnosis of depression, by no means is this the fullest extent of the vision behind it. All the recordings of putative cases will also help the medical fraternity identify and study hitherto unknown symptoms, thereby making the process even sharper.